The efficiency and eco-efficiency of manufacturing
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Gutowski, Timothy G. "The efficiency and eco-efficiency of
manufacturing." International Journal of Nanomanufacturing 6
(2010): 38 - 45.
As Published
http://dx.doi.org/10.1504/IJNM.2010.034770
Publisher
Inderscience
Version
Author's final manuscript
Accessed
Thu May 26 06:22:45 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/60894
Terms of Use
Attribution-Noncommercial-Share Alike 3.0 Unported
Detailed Terms
http://creativecommons.org/licenses/by-nc-sa/3.0/
Int. J. Nanomanufacturing, Vol. X, No. Y, xxxx
1
The efficiency and eco-efficiency of manufacturing
Timothy G. Gutowski
Massachusetts Institute of Technology,
77 Massachusetts Avenue,
Cambridge, MA 02139-4307, USA
E-mail: Gutowski@mit.edu
Abstract: In this paper, we review the efficiency of both manufacturing
processes and systems over recent decades and compare nano-materials
technologies in this context. To a first approximation, nano-materials processes
appear to be about as efficient as semi-conductor processes. That is, their
second law efficiencies are of the order 10–5, while conventional processes are
of order 10–2. However, many of these processes are early in their development
and some opportunities do exist for improvement. At the same time, many
aspects of these processes (the need for high purity materials, low yields, and
operating conditions far from equilibrium) may make these materials
vulnerable to high energy prices.
Keywords: thermodynamic efficiency; manufacturing processes; exergy;
degree of perfection; rebound.
Reference to this paper should be made as follows: Gutowski, T.G. (xxxx)
‘The efficiency and eco-efficiency of manufacturing’, Int. J.
Nanomanufacturing, Vol. X, No. Y, pp.000-000.
Biographical notes: Timothy G. Gutowski is a Professor of Mechanical
Engineering at the Massachusetts Institute of Technology.
1
Introduction
In this paper, we review the efficiency and eco-efficiency of manufacturing processes and
systems, and present data which locates nano-materials and manufacturing processes for
nano-materials in this larger context. The paper has two parts. In the first part, we review
the thermodynamic definitions of process efficiency based on the first and second laws.
Second law efficiencies, including the so-called ‘degree of perfection’, are calculated for
several different manufacturing processes ranging to nano-scale materials and devices. In
the second part of this paper, we present the idea of eco-efficiency, which is generally the
ratio of product output divided by some measure of environmental impact. For example,
manufacturing processes can be described in terms of energy consumed per unit of
material processed. Eco-efficiency would be the reciprocal of this term. Eco-efficiency
data is presented for manufacturing processes as well as industrial sectors. Again, the
performance of nano-manufacturing will be located relative to other more traditional
methods of manufacturing.
Copyright © 200x Inderscience Enterprises Ltd.
2
2
T.G. Gutowski
Thermodynamic efficiency of materials conversion processes
Efficiency, ‘η’, is usually formed as a ratio of the useful output to the inputs. The number
is dimensionless and can range from 0, (totally inefficient process) to 1, (a perfect
process). Applying this definition to a heat engine for example, one could form the
dimensionless ratio of the work out, Wout, divided by the heat input, Qin.
η=
W out
Q in
(1)
This is the conventional first law efficiency for a heat engine. While this type of measure
is commonly reported in the literature, it fails on one important dimension. It is easy to
show that no heat engine can ever obtain an efficiency of one. This shortcoming can be
corrected by applying a second law analysis to a heat engine, which states that the useful
work output cannot exceed the heat input times the so-called Carnot factor (1 – TL / TH)
where the temperatures correspond to the low and high temperature reservoirs. This is
also a dimensionless ratio and can in fact range from zero to one, with one representing a
perfect, so-called reversible heat engine.
η II =
W out
T
Q H (1 − L )
TH
(2)
In this calculation, we used the result that the maximum work output is the function of
the heat input from the high temperature reservoir QH times the Carnot factor.
If the low temperature heat is exhausted at the environmental reference temperature
say T0, the denominator in equation (2) gives the exergy content of the input heat. Exergy
is a thermodynamic variable that gives the work potential of a system with respect to a
reference environment. Exergy accounting includes the work potential of heat and work
interactions, as well as the work potential of the physical and chemical states of the
materials. It is particularly useful when analysing material transformation processes, such
as those that take place in manufacturing. Exergy can be defined as the difference in a
free energy type term in reference to an identified reference environment (denoted by the
subscript ‘o’) as given below.
B = ( H − To S ) − ( H o − To S o )
(3)
Here, H is enthalpy, S is entropy and B is exergy (see Gyftopoulos and Beretta, 2005;
de Swaan Arons et al., 2004; Szargut et al., 1988). In terms of material transformation
processes such as used in manufacturing, it can be shown that the minimum work input
required to effect the transformation is just the difference between the exergy output
minus the exergy input (Branham et al., 2008). From this formulation, we can estimate
the efficiency of a transformation process by taking the ratio of the exergy of the useful
output and divide by the exergy of the combined inputs. This ratio is called the ‘degree of
perfection’ and is denoted as ηp (Szargut et al., 1988).
The efficiency and eco-efficiency of manufacturing
ηp =
3
B out
B in
(4)
When one considers the manufacturing processes in which the materials enter and exit at
the environmental temperature and pressure, the exergy of the output is essentially the
exergy of the processed materials. The exergy of the inputs for most of the processes
considered here includes electricity and materials. Hence, the degree of perfection can be
calculated from a list of input and output materials and energy interactions for any
materials transformation process. We have calculated these values for a number of
conventional processes (i.e. sand casting and injection molding), semi conductor
processes (i.e. plasma enhanced chemical vapour deposition, sputtering and the wet
thermal oxidation process) and nano-materials processes (primarily variations on the
chemical vapour deposition process), see Table 1.
Table 1
ηp for various manufacturing processes
ηp
Processes
Injecting moulding,
Induction melting of iron
0.7 – 0.9
Semi-conductor processes
10–4 – 10–7
Nano-materials
10–3 – 10–7
To construct this table, electricity was counted as pure exergy (losses at the utility were
not counted) and the material exergy values and reference system developed by
Szargut et al. (1988) was employed. As one can see, the conventional processes have
relatively high values of ηp due to a high transit exergy (exergy embodied in the material
throughput). The semi conductor processes have very low values due to several effects;
relatively low output exergies, low material utilisation, the use of high exergy auxiliary
materials (mostly for cleaning which are destroyed as part of the abatement process), and
operation at relatively high temperatures and low pressures. The nano-materials processes
can be in the same range as semi-conductor processes or slightly higher due to a number
of possible effects including: higher output material exergies, higher yields, and more
modest temperature and pressure requirements. For example, Nanda et al. (2003) review
results for the surface energies of nano-particles.
In some manufacturing processes, the object is to remove material (milling, turning,
etching, etc.), in which case the output is in fact a hole or an area of missing material in
the incoming solid raw materials. In this case, rather than using the degree of perfection,
one might prefer an alternative definition for a second law efficiency, such as the ratio of
the minimum work required to perform the specified operation divided by the exergy of
the inputs. This is given in the equation below (see Branham et al., 2008).
η II =
W min
B in
(5)
4
3
T.G. Gutowski
The eco-efficiency of manufacturing
Eco-efficiency is defined as the ratio of a useful output divided by some measure of
environmental impact or resource use. Hence, it is most commonly not a dimensionless
ratio. An example of eco-efficiency would be the ratio of say one kilogram of aluminium
divided by the energy required to produce one kilogram of aluminium. This is, in fact, the
reciprocal of the energy intensity of aluminium, which is in the range of
100 – 200 MJ/kg. Hence, eco-efficiency is often the reciprocal of some intensity metric.
By convention, the energy referred to here is the cumulative equivalent fossil fuel
energy in terms of lower heating values. This cumulative energy value would include all
of the steps required to make the particular product under consideration. For example, if
the product is a kilogram of aluminium, then all of the energy required to make this
aluminium, including mining, milling and smelting would be accounted for. Clearly,
these values are dependent on a number of factors including typical ore grades and
technologies employed. As a result, energy intensities or eco-efficiencies depend heavily
on the boundaries and details of the study employed.
4
The eco-efficiency of manufacturing process
If one tracks the historical development of any particular manufacturing technology,
generally one finds that the eco-efficiency improves with time. That is, the resources used
per unit of production are reduced. Particular attention is focused on those resources
which are the most expensive. A particularly clear illustration of this is the data on
cutting times in machining over the course of the last hundred years. The data show that
over the last one hundred years, the cutting time for machining has been reduced by about
a factor of 100 due in large part by new developments in cutting tool materials
(Kalpakjian, 1995). Because considerable energy can be consumed by auxiliary
equipment as well as the cutting spindle, this also translates into a significant reduction in
the energy consumed per unit of material machined.
However, if one looks at the development of new manufacturing processes over the
same time period, one sees a significantly different pattern. The general trend is to
substitute low cost resources for high cost resources to gain some competitive advantage.
Since over the last several decades energy prices have been stable and declining (omitting
our most recent history) this has translated into the substitution of energy and materials
for human labour. Furthermore, recent progress in manufacturing has emphasised the
development of techniques to make micro and nano scale devices. The result has been a
very significant increase in the energy used per unit of material processed. In a recent
publication, we have reviewed 20 different manufacturing processes ranging from
conventional processes (i.e. machining, grinding, injection moulding, sand casting,
die casting), advanced machining processes (abrasive water jet, wire and drill electrical
discharge machining), rapid prototyping (laser direct metal deposition), semi conductor
additive processes (chemical vapour deposition, plasma enhanced chemical vapour
deposition, thermal oxidation and sputtering), semi conductor subtractive processes (dry
etching of oxide and nitride films) and nano-materials production processes (single
walled carbon nano tubes, carbon nano tubes and carbon nano fibres) and have found that
the energy intensity of these processes has increased by almost eight orders of
The efficiency and eco-efficiency of manufacturing
5
magnitude over the last several decades. The results are summarised in Table 2
(Gutowski et al., 2009).
Table 2
The energy intensity of manufacturing processes
Process
Approximate energy intensity (J/kg of material processes)
Conventional
1 – 20 × 106
Advanced machining
2 × 108 – 2 × 1012
Rapid prototyping
1010
Semi conductor additive
109 – 5 × 1013
Semi conductor subtractive
2 × 1010
Nano-materials
109 – 1012
5
The eco-efficiency of industrial sectors
The concept of eco-efficiency or impact/resource intensity is applicable over any size
scale of activity. We have investigated the historical eco-efficiency of ten different
production and consumption sectors in terms of energy resources used. A summary of the
sectors, time periods, and parameters measured are given in Table 3. In general, we
collected annual data on the quantity of production or consumption Q, and the resources
used R. The time periods for some of these observations are very long, in the case of
pig-iron starting in 1800 (Dahmus and Gutowski, 2009). Our discussion of this data can
be framed using a simple mathematical identity that relates resources used R to the
quantity produced or consumed Q.
R = Q.
R
1
= Q.
Q
e
(6)
In this expression, e is the so-called eco-efficiency measured in terms of quantity
produced or consumed Q divided by the resources used R. For infinitesimal changes, one
can write an expression for the incremental change in resources used in terms of the
incremental change in quantity Q and eco-efficiency e.
ΔR ΔQ Δe
=
−
R
Q
e
(7)
It is quite clear that to reduce resources, eco-efficiency must outpace production or
consumption. Our review of the data in these ten sectors shows that, in general for the
industrial sectors, eco-efficiency increases unrelentingly throughout the history of
production. However, in almost every case, the quantity produced increases faster
resulting in an increase in resources used. Hence, as a general trend, based on historical
observation, eco-efficiency improvement does not lead to resource conservation. In
reviewing the 75 decades of data represented in Table 3, we did find several exceptions,
decades of industrial activity when resources used actually decreased. In other words,
eco-efficiency outpaced production or consumption. This almost always happened due to
either
T.G. Gutowski
6
1
government intervention such as efficiency standards
2
increased prices.
The reasons for increased production or consumption can be understood by expanding
equation (6) to include additional factors.
R = P.
D Q R
. .
P D Q
(8)
In this expression P represents population, and D dollars. Hence the first term in
equation (8) accounts for increases in population and/or market size. The second term
accounts for buying power or affluence. The third term represents choice on how one
spends their dollars to buy a particular quantity of product Q, and the last term is, of
course, the reciprocal of eco-efficiency. The relentless increase in resources used, both in
production and consumption, is in large part due to our relentless growth in population
and market size as well as affluence and spending power. This equation is similar to the
so-called IPAT equation, which has a significant literature (Ehrlich and Holdren, 1972;
Commoner, 1971; Chertow, 2000).
Table 3
Production and consumption sectors studied for resource use
Activity
Dates
Boundary
Quantity
Resource
Pig-iron
1800–1990
World
kg pig iron
Joules of coke
Aluminium
1900–2005
World
kg aluminium
Joule of electricity
Nitrogen fertiliser
1915–2000
World
kg Nitrogen
Joule energy
Electricity from
coal
1920–2007
US
Joule electricity
kg coal
Electricity from oil
1920–2007
US
Joule electricity
Liter of oil
Electricity from
natural gas
1920–2007
US
Joule electricity
m3 of natural gas
Freight rail
1960–2006
US
Revenue tonne – km
Liter fuel
Air travel
1960–2005
US
Seat – km
Liter fuel
Motor vehicle
1936–2006
US
Vehicle – km
Liter fuel
Refrigeration
1960–2002
US
Hours refrigeration
Joule electricity
In this discussion, it is important to point out that the terms in equations (6), (7) and (8)
may not be independent. That is, affluence affects population growth, quantity produced
affects eco-efficiency and so forth. For this discussion, we are particularly interested in
the interaction between the quantity Q and the efficiency e represented in equations (6)
and (7). There are several well known effects that cause these variables to interact. For
example, learning affects and economies of scale, suggest that as the scale of production
Q increases, this leads to an increase in the efficiency of production. Hence large Q leads
to increased eco-efficiency. On the other hand, it has also been observed that increased
efficiency can result in increased demand and larger Q. This is so-called the rebound
effect. In simple terms, rebound can occur because efficiency improvements can lead to
price reductions. In the case where there is a significant elasticity of demand, price
reductions can in turn lead to increased quantities produced and consumed. For example,
we could capture the rebound effect by writing
The efficiency and eco-efficiency of manufacturing
ΔQ / Q ΔQ / Q Δπ / π
⋅
= ε ⋅γ
=
Δπ / π Δe / e
Δe / e
7
(9)
In this expression, ε is the so-called own price elasticity. The second term captures how
price, π, is affected by changes in eco-efficiency. In general ε may vary from very low
values for in-elastic goods, to numbers much larger than one for highly elastic goods
(see for example, Gwartney et al., 2000, p.510, for elasticities that range from 0.1 to 4.6).
The magnitude of γ depends on what fraction of the product cost is represented by the
resource, how large the profit is, and to what extent the efficiency increase is converted
into a price reduction. In any case 0 < ⎜ γ ⎜ < 1, and strictly speaking both ε and γ are
actually negative. The effect of the rebound can be illustrated by first considering the
case where there is an improvement in eco-efficiency but no change in quantity, then
from equation (7) one gets a clear reduction in resource use, i.e. ∆R/R = –∆e/e. However,
if there is a rebound effect and the quantity Q increases according to equation (9), then
substitution into equation (7) gives
ΔR
Δe
= ( ε ⋅γ −1)
R
e
(10)
In the remote (but not impossible) case where ε ⋅ γ = 1 , the entire savings due to
improved eco-efficiency could be negated by the rebound effect. Numerous
micro-economic studies show that ε ⋅ γ is finite but less than one (Greening et al., 2000;
Kalakkad Jayaraman, 2008). At the macro economic level, some have put forth the
argument that it is in fact our very efficiency that drives growth (Herring, 1998). While
this is a very contentious debate, it is worth paying attention to it, as our future may
depend upon us understanding these mechanisms.
References
Branham, M., Gutowski, T., Jones, A. and Sekulic, D. (2008) ‘A thermodynamic framework for
analyzing and improving manufacturing processes’, IEEE International Symposium on
Electronics and the Environment, 19–20 May, San Francisco, USA.
Chertow, M. (2000) ‘The IPAT equation and its variants: changing views of technology and
environmental impact’, Journal of Industrial Ecology, Vol. 4, No. 4, p.13–29.
Commoner, B. (1971) The Closing Circle, Knopf, Inc., USA.
Dahmus, J.B., and Gutowski, T. (2009) ‘Can efficiency improvements reduce resource
consumption? A historical analysis of ten activities’, accepted for publication to Journal of
Industrial Ecology.
de Swaan Arons, J., van der Kooi, H. and Sankaranarayanan, K. (2004) Efficiency and
Sustainability in the Energy and Chemical Industries, Marcel Dekker Inc., New York.
Ehrlich, P.R. and Holdren, J.P. (1972) ‘Critique: one dimensional ecology’, Bulletin of the Atomic
Scientists, Vol. 28, No. 5, pp.18–27.
Greening, L.A., Greene, D.L. and Difiglio, C. (2000) ‘Energy efficiency and consumption – the
rebound effect – a survey’, Energy Policy, Vol. 28, Nos. 6–7, pp.389–401.
Gutowski, T.G., Branham, M.S., Dahmus, J.B., Jones, A.J., Thiriez, A. and Sekulic, D.. (2009) ‘A
thermodynamic analysis of resources used in manufacturing processes’, Environmental
Science and Technology, Vol. 43, pp.1584–1590.
8
T.G. Gutowski
Gwartney, J.D., Stroup, R.L. and Sobel, R.S. (2000) Economics – Private and Public Choice, 9th
ed., South-Western, Mason, Ohio.
Gyftopoulos, E. and Beretta, G. (2005) Thermodynamics – Foundations and Applications, Dover,
Mineola, New York.
Herring, H. (1998) ‘Does energy efficiency save energy: the economists debate’, Energy and
Environment Research Unit Report No. 074, July, available online at
http://technology.open.ac.uk/eeru/staff/horace/hh3.htm.
Kalakkad Jayaraman, S.K. (2008) ‘Computation of economic rebound effect in different sectors of
the U.S. economy’, S.M. Thesis, Massachusetts Institute of Technology, Department of
Mechanical Engineering.
Kalpakjian, S. (1995) Manufacturing Engineering and Technology, 3rd ed., Addison-Wesley
Publishing Company, Reading, Massachusetts.
Nanda, K.K., Maisels, A., Kruis, F.E., Fissan, H. and Stappert, S. (2003) ‘Higher surface energy of
free nanoparticles’, Physical Review Letters, The American Physical Society, Vol. 91,
No. 10, pp.106102–4.
Szargut, J., Morris, D.R. and F.R. Steward (1988) Energy Analysis of Thermal Chemical and
Metallurgical Processes, Hemisphere Publishing Corporation and Springer-Verlag, New
York.